Unauthorised AI agents are reportedly already accessing US bank accounts

[Freya]

AI agents are turning up in bank accounts without the banks' approval. Engadget, citing reporting from Axios, says unauthorised AI agents are already accessing customer accounts at major US banks

The banks named include Morgan Stanley, Bank of America, JPMorgan and Wells Fargo. Customers are giving AI agents their login details so the agents can check balances, move money or handle tasks for them. The banks never agreed to this kind of access

JPMorgan's Jamie Dimon has warned about the possibility of thousands of rogue agents. Banks worry that they cannot tell a customer's own agent from a criminal one, and that agents could make mistakes with real money

There is a legal question too. The Electronic Fund Transfer Act protects consumers from unauthorised transactions, but if a customer handed their credentials to an agent, losses may not be covered. People could be left out of pocket if something goes wrong

This feels like the early days of screen scraping all over again, before banks built proper APIs for apps. The sensible fix is official, limited access for agents, rather than sharing passwords. Until then, giving an AI your bank login is a big risk

Would you let an AI agent into your bank account? And should banks build official access for agents?

Unauthorized AI agents are reportedly already accessing US bank accounts - Engadgetwww.engadget.com